The goal is to produce radiolabeled monoclonal antibodies (MAb) to CEA or GICA useful in the diagnosis, anatomic localization and management of human colorectal cancer using immunologic and radionuclide methods which already have been established. The MAbs will be obtained from Mach in Switzerland, and Koprowski at the Wistar Institute, and they will be labeled with 131I or 123I by the iodine monochloride technique or the iodobead method, or they will be labeled with 111In, 88Y or 90Y after coupling with the cyclic dianhydride of DTPA. These labeling methods have not interfered with the specific binding to human colon cancer cell lines in vitro. The biodistribution of these labeled MAbs in normal mice has been studied already. The tumor localization and organ biodistribution of these labeled MAbs will be determined in nude mice bearing subcutaneous or intrahepatic transplants of two human colon cancer cell lines CO-112 and LS174T using tissue sampling and counting or external scintigraphy. The use of an antigen-negative tumor (Raji, CEM or Namalwa) in nude mice will be used to control for in vivo specificity. Labeled antibody fragments will also be studied in this manner. The parameters to be investigated are the antibody, radiolabel, dose, time after administration and tumor size. Therapeutic studies will be conducted in nude mice with human colon cancer xenografts and 131I or 90Y beta-emitting radio-nuclides conjugated to the MAbs in an effort to determine the potential use of beta emitters for radioimmunotherapy. The dependency of successful colon carcinoma localization on the circulating CEA or GICA antigen level will also be determined. The immunohistochemical binding of MAb directed against CEA or GICA to biopsies of colorectal cancer and normal tissues will be studied in vitro. In the third year, a pilot clinical study will be initiated evaluating one or more of the best radiolabeled monoclonal antibody preparations, based on the results of the tissue binding and animal localization studies.